摘要
为实现二级倒立摆系统的实时稳定控制,以深圳固高直线二级倒立摆装置作为控制对象,在MATLAB环境下,利用基于二次型最优控制理论的线性二次型(Linear Quadratic Regulator,LQR)最优控制器,成功实现了该装置的实时稳定控制。为引入新的控制策略,采集二级倒立摆实时控制过程中的LQR控制器数据作为样本,经过自适应神经模糊推理系统(Adaptive Neuro-Fuzzy Inference System,ANFIS)工具箱训练并生成出一种新型模糊神经网络控制器,应用到装置上同样实现了实时平衡。结果表明,新型控制器较LQR控制器控制效果更优,也为成功实现装置的实时平衡提供了一种新的思路和解决方法。
To realize real - time equilibrium control of double - link inverted pendulum system, a linear double - link inverted pendulum device of Googol Tech is taken as the controlling object. In the environment of MATLAB, a linear quadratic regulator （LQR） controller based on quadratic optimal control theory was applied to the device. And real -time equilibrium control of the system succeeded. In order to introduce new control strategy, the data of the LQR controller were collected during the real - time controlling experiment of the LQR controller and trained in adaptive neuro- fuzzy inference system （ANFIS） toolbox of MATLAB as a sample. A new fuzzy- neural -network controller was generated and the real - time balance of the device was also achieved by using it. Compared with the LQR controller, a better controlling effect was gained, and it also provides a new thought and solution for realization of the experiment.
出处
《计算机仿真》
CSCD
北大核心
2009年第9期342-345,共4页
Computer Simulation
基金
北京市教委科技计划面上项目(KM200611232012)
关键词
二级倒立摆
线性二次型
模糊神经网络
自适应神经模糊推理系统
实时稳定控制关键词:
Double - link inverted pendulum
Linear quadratic regulator （LQR）
Fuzzy - neural - network
A- daptive neuro- fuzzy inference system
Real -time equilibrium control
